Message Boards Message Boards

0
|
3990 Views
|
3 Replies
|
0 Total Likes
View groups...
Share
Share this post:

Imageidentify on Rapsberry PI

Posted 9 years ago

Hello everybody, I'm new to this topic, although i've used the Wolfram alpha webpage on various occasions. I saw an article on Gizmodo about Image identification using the wolfram engine (the webpage) and I thought I will use it for a school project. What I am trying to achieve is to use a rapsberry pi (mounted on an Rc car) to identify the world and map it (it also has a gps). I've installed the wolfram engine and tried to identify the objects taken with the camera. I've used the documentation here: https://reference.wolfram.com/language/ref/ImageIdentify.html As I have only the CLI interface, I cannot see the outputs and inputs graphically. (Being a project that has to be mobile, I don't want to be running a gui.) This is what I've tried:

In[3]:= camImage=Import[ "!raspistill -ss 1000 -t 10 -w 1024 -h 1024 -o -", "JPG"]

Out[3]= -Image-

In[4]:= ImageIdentify[camImage]

Out[4]= ImageIdentify[-Image-]

I've expected that on the output, I receive the textual description of the object, instead I am only getting the string above (Out[4]). I've also tried using a jpg stored locally, but with the same results. Note:! As this is run in a command line interface, there is no way to see any graphical image...

Am I doing something wrong ?
Thanks in advance!

POSTED BY: Barna Peto
3 Replies
Posted 9 years ago

So in the end I've managed to implement the project by using a linux machine that had version 10.1 . The raspberry connects wia ssh and executes a wolfram script and returns the result to the host machine.

POSTED BY: Barna Peto
Posted 9 years ago

Oh.. I see. Thanks for the answer. I didn't thought about that. I know this is a standard question, but is there an estimation when will the new release available on the Raspberry platform ? I'm asking this, as there is a deadline when I have to finish the project.

Another thought: I could use curl and some other script combinations to upload an image to the www.imageidentify.com website, and parse the result. I saw however that there is a javascript doing the upload job. I'm only familiar with POST-ing the jpg file... I will have to research a bit, but if anybody here has knowledge in how to do it, then any help is greatly appreciated!!

POSTED BY: Barna Peto

ImageIdentify is new in version 10.1, while the latest available Raspberry Pi pilot release build is 10.0.2.

POSTED BY: Ilian Gachevski
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract